static UBool
do_canonicalize(const char* localeID,
char* buffer,
int32_t bufferCapacity,
UErrorCode* err)
{
uloc_canonicalize(
localeID,
buffer,
bufferCapacity,
err);
if (*err == U_STRING_NOT_TERMINATED_WARNING ||
*err == U_BUFFER_OVERFLOW_ERROR) {
*err = U_ILLEGAL_ARGUMENT_ERROR;
return FALSE;
}
else if (U_FAILURE(*err)) {
return FALSE;
}
else {
return TRUE;
}
}
int32 FInternationalization::GetCultureIndex(const FString& Name)
{
#if UE_ENABLE_ICU
static const int32 MaximumNameLength = 64;
const int32 NameLength = Name.Len();
check(NameLength < MaximumNameLength);
char CanonicalName[MaximumNameLength];
UErrorCode ICUStatus = U_ZERO_ERROR;
uloc_canonicalize(TCHAR_TO_ANSI( *Name ), CanonicalName, MaximumNameLength, &ICUStatus);
#endif
const int32 CultureCount = AllCultures.Num();
int32 i;
for (i = 0; i < CultureCount; ++i)
{
#if UE_ENABLE_ICU
if( AllCultures[i]->GetName() == CanonicalName )
#else
if( AllCultures[i]->GetName() == Name )
#endif
{
break;
}
}
if(i >= CultureCount)
{
i = -1;
}
return i;
}
U_CDECL_END
U_NAMESPACE_BEGIN
Locale *locale_set_default_internal(const char *id, UErrorCode& status) {
// Synchronize this entire function.
Mutex lock(&gDefaultLocaleMutex);
UBool canonicalize = FALSE;
// If given a NULL string for the locale id, grab the default
// name from the system.
// (Different from most other locale APIs, where a null name means use
// the current ICU default locale.)
if (id == NULL) {
id = uprv_getDefaultLocaleID(); // This function not thread safe? TODO: verify.
canonicalize = TRUE; // always canonicalize host ID
}
char localeNameBuf[512];
if (canonicalize) {
uloc_canonicalize(id, localeNameBuf, sizeof(localeNameBuf)-1, &status);
} else {
uloc_getName(id, localeNameBuf, sizeof(localeNameBuf)-1, &status);
}
localeNameBuf[sizeof(localeNameBuf)-1] = 0; // Force null termination in event of
// a long name filling the buffer.
// (long names are truncated.)
//
if (U_FAILURE(status)) {
return gDefaultLocale;
}
if (gDefaultLocalesHashT == NULL) {
gDefaultLocalesHashT = uhash_open(uhash_hashChars, uhash_compareChars, NULL, &status);
if (U_FAILURE(status)) {
return gDefaultLocale;
}
uhash_setValueDeleter(gDefaultLocalesHashT, deleteLocale);
ucln_common_registerCleanup(UCLN_COMMON_LOCALE, locale_cleanup);
}
Locale *newDefault = (Locale *)uhash_get(gDefaultLocalesHashT, localeNameBuf);
if (newDefault == NULL) {
newDefault = new Locale(Locale::eBOGUS);
if (newDefault == NULL) {
status = U_MEMORY_ALLOCATION_ERROR;
return gDefaultLocale;
}
newDefault->init(localeNameBuf, FALSE);
uhash_put(gDefaultLocalesHashT, (char*) newDefault->getName(), newDefault, &status);
if (U_FAILURE(status)) {
return gDefaultLocale;
}
}
gDefaultLocale = newDefault;
return gDefaultLocale;
}
U_CAPI void U_EXPORT2
ucol_prepareShortStringOpen( const char *definition,
UBool,
UParseError *parseError,
UErrorCode *status)
{
if(U_FAILURE(*status)) return;
UParseError internalParseError;
if(!parseError) {
parseError = &internalParseError;
}
parseError->line = 0;
parseError->offset = 0;
parseError->preContext[0] = 0;
parseError->postContext[0] = 0;
// first we want to pick stuff out of short string.
// we'll end up with an UCA version, locale and a bunch of
// settings
// analyse the string in order to get everything we need.
CollatorSpec s;
ucol_sit_initCollatorSpecs(&s);
ucol_sit_readSpecs(&s, definition, parseError, status);
ucol_sit_calculateWholeLocale(&s);
char buffer[internalBufferSize];
uprv_memset(buffer, 0, internalBufferSize);
uloc_canonicalize(s.locale, buffer, internalBufferSize, status);
UResourceBundle *b = ures_open(U_ICUDATA_COLL, buffer, status);
/* we try to find stuff from keyword */
UResourceBundle *collations = ures_getByKey(b, "collations", NULL, status);
UResourceBundle *collElem = NULL;
char keyBuffer[256];
// if there is a keyword, we pick it up and try to get elements
if(!uloc_getKeywordValue(buffer, "collation", keyBuffer, 256, status)) {
// no keyword. we try to find the default setting, which will give us the keyword value
UResourceBundle *defaultColl = ures_getByKeyWithFallback(collations, "default", NULL, status);
if(U_SUCCESS(*status)) {
int32_t defaultKeyLen = 0;
const UChar *defaultKey = ures_getString(defaultColl, &defaultKeyLen, status);
u_UCharsToChars(defaultKey, keyBuffer, defaultKeyLen);
keyBuffer[defaultKeyLen] = 0;
} else {
*status = U_INTERNAL_PROGRAM_ERROR;
return;
}
ures_close(defaultColl);
}
collElem = ures_getByKeyWithFallback(collations, keyBuffer, collElem, status);
ures_close(collElem);
ures_close(collations);
ures_close(b);
}
FString FCulture::FICUCultureImplementation::GetCanonicalName(const FString& Name)
{
static const int32 MaximumNameLength = 64;
const int32 NameLength = Name.Len();
check(NameLength < MaximumNameLength);
char CanonicalName[MaximumNameLength];
UErrorCode ICUStatus = U_ZERO_ERROR;
uloc_canonicalize(TCHAR_TO_ANSI( *Name ), CanonicalName, MaximumNameLength, &ICUStatus);
return CanonicalName;
}
int32_t GetLocale(
const UChar* localeName, char* localeNameResult, int32_t localeNameResultLength, bool canonicalize, UErrorCode* err)
{
char localeNameTemp[ULOC_FULLNAME_CAPACITY] = {0};
int32_t localeLength;
// Convert ourselves instead of doing u_UCharsToChars as that function considers '@' a variant and stops.
for (int i = 0; i < ULOC_FULLNAME_CAPACITY - 1; i++)
{
UChar c = localeName[i];
if (c > (UChar)0x7F)
{
*err = U_ILLEGAL_ARGUMENT_ERROR;
return ULOC_FULLNAME_CAPACITY;
}
localeNameTemp[i] = (char)c;
if (c == (UChar)0x0)
{
break;
}
}
if (canonicalize)
{
localeLength = uloc_canonicalize(localeNameTemp, localeNameResult, localeNameResultLength, err);
}
else
{
localeLength = uloc_getName(localeNameTemp, localeNameResult, localeNameResultLength, err);
}
if (U_SUCCESS(*err))
{
// Make sure the "language" part of the locale is reasonable (i.e. we can fetch it and it is within range).
// This mimics how the C++ ICU API determines if a locale is "bogus" or not.
char language[ULOC_LANG_CAPACITY];
uloc_getLanguage(localeNameTemp, language, ULOC_LANG_CAPACITY, err);
if (*err == U_STRING_NOT_TERMINATED_WARNING)
{
// ULOC_LANG_CAPACITY includes the null terminator, so if we couldn't extract the language with the null
// terminator, the language must be invalid.
*err = U_ILLEGAL_ARGUMENT_ERROR;
}
}
return localeLength;
}
static ULayoutType
_uloc_getOrientationHelper(const char* localeId,
const char* key,
UErrorCode *status)
{
ULayoutType result = ULOC_LAYOUT_UNKNOWN;
if (!U_FAILURE(*status)) {
int32_t length = 0;
char localeBuffer[ULOC_FULLNAME_CAPACITY];
uloc_canonicalize(localeId, localeBuffer, sizeof(localeBuffer), status);
if (!U_FAILURE(*status)) {
const UChar* const value =
uloc_getTableStringWithFallback(
NULL,
localeBuffer,
"layout",
NULL,
key,
&length,
status);
if (!U_FAILURE(*status) && length != 0) {
switch(value[0])
{
case 0x0062: /* 'b' */
result = ULOC_LAYOUT_BTT;
break;
case 0x006C: /* 'l' */
result = ULOC_LAYOUT_LTR;
break;
case 0x0072: /* 'r' */
result = ULOC_LAYOUT_RTL;
break;
case 0x0074: /* 't' */
result = ULOC_LAYOUT_TTB;
break;
default:
*status = U_INTERNAL_PROGRAM_ERROR;
break;
}
}
}
}
return result;
}
static void php_canonicalize_locale_id(char **target, int32_t *target_len, char *locale, UErrorCode *status)
{
char *canonicalized = NULL;
int32_t canonicalized_len = 128;
while (1) {
*status = U_ZERO_ERROR;
canonicalized = erealloc(canonicalized, canonicalized_len + 1);
canonicalized_len = uloc_canonicalize(locale, canonicalized, canonicalized_len, status);
if (*status != U_BUFFER_OVERFLOW_ERROR) {
break;
}
}
canonicalized[canonicalized_len] = 0;
*target = canonicalized;
*target_len = canonicalized_len;
}
/* {{{
* Gets the value from ICU
* common code shared by get_primary_language,get_script or get_region or get_variant
* result = 0 if error, 1 if successful , -1 if no value
*/
static char* get_icu_value_internal( char* loc_name , char* tag_name, int* result , int fromParseLocale)
{
char* tag_value = NULL;
int32_t tag_value_len = 512;
int singletonPos = 0;
char* mod_loc_name = NULL;
int grOffset = 0;
int32_t buflen = 512;
UErrorCode status = U_ZERO_ERROR;
if( tag_name != LOC_CANONICALIZE_TAG ){
/* Handle grandfathered languages */
grOffset = findOffset( LOC_GRANDFATHERED , loc_name );
if( grOffset >= 0 ){
if( strcmp(tag_name , LOC_LANG_TAG)==0 ){
tag_value = estrdup(loc_name);
return tag_value;
} else {
/* Since Grandfathered , no value , do nothing , retutn NULL */
return NULL;
}
}
if( fromParseLocale==1 ){
/* Handle singletons */
if( strcmp(tag_name , LOC_LANG_TAG)==0 ){
if( strlen(loc_name)>1 && (isIDPrefix(loc_name) ==1 ) ){
return loc_name;
}
}
singletonPos = getSingletonPos( loc_name );
if( singletonPos == 0){
/* singleton at start of script, region , variant etc.
* or invalid singleton at start of language */
return NULL;
} else if( singletonPos > 0 ){
/* singleton at some position except at start
* strip off the singleton and rest of the loc_name */
mod_loc_name = estrndup ( loc_name , singletonPos-1);
}
} /* end of if fromParse */
} /* end of if != LOC_CANONICAL_TAG */
if( mod_loc_name == NULL){
mod_loc_name = estrdup(loc_name );
}
/* Proceed to ICU */
do{
tag_value = erealloc( tag_value , buflen );
tag_value_len = buflen;
if( strcmp(tag_name , LOC_SCRIPT_TAG)==0 ){
buflen = uloc_getScript ( mod_loc_name ,tag_value , tag_value_len , &status);
}
if( strcmp(tag_name , LOC_LANG_TAG )==0 ){
buflen = uloc_getLanguage ( mod_loc_name ,tag_value , tag_value_len , &status);
}
if( strcmp(tag_name , LOC_REGION_TAG)==0 ){
buflen = uloc_getCountry ( mod_loc_name ,tag_value , tag_value_len , &status);
}
if( strcmp(tag_name , LOC_VARIANT_TAG)==0 ){
buflen = uloc_getVariant ( mod_loc_name ,tag_value , tag_value_len , &status);
}
if( strcmp(tag_name , LOC_CANONICALIZE_TAG)==0 ){
buflen = uloc_canonicalize ( mod_loc_name ,tag_value , tag_value_len , &status);
}
if( U_FAILURE( status ) ) {
if( status == U_BUFFER_OVERFLOW_ERROR ) {
status = U_ZERO_ERROR;
continue;
}
/* Error in retriving data */
*result = 0;
if( tag_value ){
efree( tag_value );
}
if( mod_loc_name ){
efree( mod_loc_name);
}
return NULL;
}
} while( buflen > tag_value_len );
if( buflen ==0 ){
/* No value found */
*result = -1;
if( tag_value ){
efree( tag_value );
}
if( mod_loc_name ){
efree( mod_loc_name);
}
return NULL;
} else {
*result = 1;
}
if( mod_loc_name ){
efree( mod_loc_name);
}
return tag_value;
}
/**
* Test localized currency patterns for PREEURO variants.
*/
static void TestCurrencyPreEuro(void)
{
UNumberFormat *currencyFmt;
UChar *str=NULL, *res=NULL;
int32_t lneed, i;
UFieldPosition pos;
UErrorCode status = U_ZERO_ERROR;
const char* locale[]={
"ca_ES_PREEURO", "de_LU_PREEURO", "en_IE_PREEURO", "fi_FI_PREEURO", "fr_LU_PREEURO", "it_IT_PREEURO",
"pt_PT_PREEURO", "de_AT_PREEURO", "el_GR_PREEURO", "es_ES_PREEURO", "fr_BE_PREEURO", "ga_IE_PREEURO",
"nl_BE_PREEURO", "de_DE_PREEURO", "en_BE_PREEURO", "eu_ES_PREEURO", "fr_FR_PREEURO", "gl_ES_PREEURO",
"nl_NL_PREEURO",
};
const char* result[]={
"\\u20A7\\u00A02", "2\\u00A0F", "IEP\\u00A01.50", "1,50\\u00A0mk", "2\\u00A0F", "ITL\\u00A02",
"1$50\\u00A0\\u200B", "\\u00F6S\\u00A01,50", "1,50\\u00A0\\u0394\\u03C1\\u03C7", "2\\u00A0\\u20A7", "1,50\\u00A0FB", "IEP\\u00A01.50",
"BEF\\u00A01,50", "1,50\\u00A0DM", "1,50\\u00A0BEF", "\\u20A7\\u00A02", "1,50\\u00A0F", "2\\u00A0\\u20A7",
"NLG\\u00A01,50"
};
log_verbose("\nTesting the number format with different currency patterns\n");
for(i=0; i < 19; i++)
{
char curID[256] = {0};
uloc_canonicalize(locale[i], curID, 256, &status);
if(U_FAILURE(status)){
log_data_err("Could not canonicalize %s. Error: %s (Are you missing data?)\n", locale[i], u_errorName(status));
continue;
}
currencyFmt = unum_open(UNUM_CURRENCY, NULL,0,curID,NULL, &status);
if(U_FAILURE(status)){
log_data_err("Error in the construction of number format with style currency: %s (Are you missing data?)\n",
myErrorName(status));
} else {
lneed=0;
lneed= unum_formatDouble(currencyFmt, 1.50, NULL, lneed, NULL, &status);
if(status==U_BUFFER_OVERFLOW_ERROR){
status=U_ZERO_ERROR;
str=(UChar*)malloc(sizeof(UChar) * (lneed+1) );
pos.field = 0;
unum_formatDouble(currencyFmt, 1.50, str, lneed+1, &pos, &status);
}
if(U_FAILURE(status)) {
log_err("Error in formatting using unum_formatDouble(.....): %s\n", myErrorName(status) );
} else {
res=(UChar*)malloc(sizeof(UChar) * (strlen(result[i])+1) );
u_unescape(result[i],res,(int32_t)(strlen(result[i])+1));
if (u_strcmp(str, res) != 0){
log_err("FAIL: Expected %s Got: %s for locale: %s\n", result[i],aescstrdup(str, -1),locale[i]);
}
}
}
unum_close(currencyFmt);
free(str);
free(res);
}
}
U_CAPI UCollator* U_EXPORT2
ucol_openFromShortString( const char *definition,
UBool forceDefaults,
UParseError *parseError,
UErrorCode *status)
{
UTRACE_ENTRY_OC(UTRACE_UCOL_OPEN_FROM_SHORT_STRING);
UTRACE_DATA1(UTRACE_INFO, "short string = \"%s\"", definition);
if(U_FAILURE(*status)) return 0;
UParseError internalParseError;
if(!parseError) {
parseError = &internalParseError;
}
parseError->line = 0;
parseError->offset = 0;
parseError->preContext[0] = 0;
parseError->postContext[0] = 0;
// first we want to pick stuff out of short string.
// we'll end up with an UCA version, locale and a bunch of
// settings
// analyse the string in order to get everything we need.
const char *string = definition;
CollatorSpec s;
ucol_sit_initCollatorSpecs(&s);
string = ucol_sit_readSpecs(&s, definition, parseError, status);
ucol_sit_calculateWholeLocale(&s);
char buffer[internalBufferSize];
uprv_memset(buffer, 0, internalBufferSize);
uloc_canonicalize(s.locale, buffer, internalBufferSize, status);
UCollator *result = ucol_open(buffer, status);
int32_t i = 0;
for(i = 0; i < UCOL_ATTRIBUTE_COUNT; i++) {
if(s.options[i] != UCOL_DEFAULT) {
if(forceDefaults || ucol_getAttribute(result, (UColAttribute)i, status) != s.options[i]) {
ucol_setAttribute(result, (UColAttribute)i, s.options[i], status);
}
if(U_FAILURE(*status)) {
parseError->offset = (int32_t)(string - definition);
ucol_close(result);
return NULL;
}
}
}
if(s.variableTopSet) {
if(s.variableTopString[0]) {
ucol_setVariableTop(result, s.variableTopString, s.variableTopStringLen, status);
} else { // we set by value, using 'B'
ucol_restoreVariableTop(result, s.variableTopValue, status);
}
}
if(U_FAILURE(*status)) { // here it can only be a bogus value
ucol_close(result);
result = NULL;
}
UTRACE_EXIT_PTR_STATUS(result, *status);
return result;
}
U_CDECL_END
U_NAMESPACE_BEGIN
UnicodeString &
LocaleUtility::canonicalLocaleString(const UnicodeString * id, UnicodeString & result)
{
if (id == NULL)
{
result.setToBogus();
}
else
{
// Fix case only (no other changes) up to the first '@' or '.' or
// end of string, whichever comes first. In 3.0 I changed this to
// stop at first '@' or '.'. It used to run out to the end of
// string. My fix makes the tests pass but is probably
// structurally incorrect. See below. [alan 3.0]
// TODO: Doug, you might want to revise this...
result = *id;
int32_t i = 0;
int32_t end = result.indexOf(AT_SIGN_CHAR);
int32_t n = result.indexOf(PERIOD_CHAR);
if (n >= 0 && n < end)
{
end = n;
}
if (end < 0)
{
end = result.length();
}
n = result.indexOf(UNDERSCORE_CHAR);
if (n < 0)
{
n = end;
}
for (; i < n; ++i)
{
UChar c = result.charAt(i);
if (c >= 0x0041 && c <= 0x005a)
{
c += 0x20;
result.setCharAt(i, c);
}
}
for (n = end; i < n; ++i)
{
UChar c = result.charAt(i);
if (c >= 0x0061 && c <= 0x007a)
{
c -= 0x20;
result.setCharAt(i, c);
}
}
}
return result;
#if 0
// This code does a proper full level 2 canonicalization of id.
// It's nasty to go from UChar to char to char to UChar -- but
// that's what you have to do to use the uloc_canonicalize
// function on UnicodeStrings.
// I ended up doing the alternate fix (see above) not for
// performance reasons, although performance will certainly be
// better, but because doing a full level 2 canonicalization
// causes some tests to fail. [alan 3.0]
// TODO: Doug, you might want to revisit this...
result.setToBogus();
if (id != 0)
{
int32_t buflen = id->length() + 8; // space for NUL
char * buf = (char *) uprv_malloc(buflen);
char * canon = (buf == 0) ? 0 : (char *) uprv_malloc(buflen);
if (buf != 0 && canon != 0)
{
U_ASSERT(id->extract(0, INT32_MAX, buf, buflen) < buflen);
UErrorCode ec = U_ZERO_ERROR;
uloc_canonicalize(buf, canon, buflen, &ec);
if (U_SUCCESS(ec))
{
result = UnicodeString(canon);
}
}
uprv_free(buf);
uprv_free(canon);
}
return result;
#endif
}
